i = 0
rf.results <- list()
  i = i + 1
  # Fit LASSO
  rf.results[[i]] <- list()

  hours <- floor((proc.time()[3]-start.time)/3600)
  mins <- floor((proc.time()[3]-start.time)/60) - hours*60
  secs <- floor((proc.time()[3]-start.time)) - hours*3600 - mins*60
  
  print(paste(hours,
              "h",
              mins,
              "m",
              secs,
              "s elapsed",
              sep = ""))
  set.seed(i)
  rf.results[[i]]$mod <-  randomForest(x = as.matrix(dta.train[,
                                                              rhs]),
                           y = as.matrix(dta.train[,
                                                  dv]),
                           ntree = 100,
                           maxnodes = 60,
                           type = regression,
                           importance = TRUE
  )
  rf.results[[i]]$fit.oos <- predict(rf.results[[i]]$mod,
                                       newdata = as.matrix(dta.test[,
                                                                    rhs]))
  rf.results[[i]]$actual.oos <- as.matrix(dta.test[,
                                                      dv])

  save(rf.results,file=paste(modeldir,"/",
                             country,
                             "_rf_",
                             v,
                             "_",
                             rhs.group,
                             "_cross",
                             fileext,
                             ".RData",
                             sep = ""))